Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
9 CHAPTER 3. RESEARCH PLAN The research plan included six components: (a) full-scale laboratory testing of typical asphalt and concrete pavement sections in an instrumented LST; (b) laboratory triaxial testing of different base courses with geosynthetics at different locations within the test samples; (c) finite element computations to match the results of the full-scale tests; (d) use of the same finite element program to develop full factorial sets of pavement data to construct ANN models of the critical strains and stresses in pavements; (e) generation of a new model of permanent deformation to predict pavement performance; and (f) comparison of the predicted performance of pavements with and without geosynthetics embedded in the unbound base courses. Full-Scale Laboratory Testing The full-scale laboratory testing plan was designed to capture the effect of two types of geosynthetics on the response of both rigid and flexible pavement sections under static and dynamic loads. Two distinctive mechanisms by which geosynthetics affected the performance of pavements were reported: (a) stiffening the entire base course, and (b) more efficiently spreading the tire load. The two geosynthetics were selected to illustrate these two separate mechanisms on both asphalt and concrete pavements. According to the literature, the benefits of geosynthetics in asphalt pavements have depended on the thickness of the crushed aggregate base (CAB) layer and the location of the geosynthetic within that layer (13, 37â41). Generally, geosynthetics were reported to be more effective in flexible pavements when placed at the base-subgrade interface of thin base sections (such as 6 inches) and near the midpoint of thicker base layers (such as 10 or more inches). Thus, two different base thicknesses (6 and 10 inches) along with two different locations for the geosynthetic were included in the test matrix for flexible pavement. Table 3.1 provides a summary of the LST experiments conducted on the flexible pavement structure. A second test matrix was developed to capture the influence of geosynthetics on rigid (Portland cement concrete [PCC]) pavement responses. For the rigid pavement experiments, the thicknesses of the PCC layer and CAB layer were kept at 6 and 8 inches, respectively. The testing plan for PCC included only a typical base thickness since the PCC slab itself supplies most of the pavementâs structural capacity. Table 3.2 provides a summary of the LST experiments conducted on the rigid pavement structure. Generally, two types of loading with three different intensities were applied to each pavement structure. The flexible pavement structure was subjected to repeated dynamic loads of 9, 12, and 16 kips for different cycles, with a pulse duration of 0.1 second followed by a rest period of 0.9 second in each loading cycle. The pavement structure was then allowed to recover under 100-lb seating load for 300 seconds. A static load of 6, 9, and 12 kip was then applied for 300 seconds with a rest period between two load levels of 30 minutes. The rigid pavement structure was subjected to the same dynamic load but only for 25 cycles with a 180-second recovery period. The static load was almost similar but was applied for 180 seconds with a rest period of 300 seconds. The loading and rest periods were longer with the flexible pavement structure to permit complete viscoelastic recovery. Table 3.3 provides a summary of the loading protocol for all of the flexible and rigid pavement experiments. Following each loading level, a
10 rest period was imposed to allow the pavement to rebound from the applied loading strains accumulated in the previous test. All loads were applied through a loading plate to the surface of the pavement layer. For the flexible pavement structure, the loading plate was located at the center of the LST, while for the rigid pavement structure, the plate was placed at the edge of the concrete slab. The measurements made in all of the test pavements included the surface deflections, vertical and horizontal strains and stresses, and relative displacement between the geosynthetic and the surrounding unbound base course. Table 3.1. LST Experiment Design for Flexible Pavement Experiment Asphalt Layer Thickness (inch) CAB Layer Thickness (inch) Reinforcement ID No. Type Location AC-Contr-B06 1 6 6 None (Control) N/A AC-Contr-B10 2 6 10 None (Control) N/A AC-Grid-B06 3 6 6 Geogrid Base-Subgrade Interface AC-Grid-B10 4 6 10 Geogrid Middle of the Base AC-Textile-B06 5 6 6 Geotextile Base-Subgrade Interface AC-Textile-B10 6 6 10 Geotextile Middle of the Base Note: AC = asphalt concrete; N/A = not applicable. Table 3.2. LST Experiment Design for Rigid Pavement Experiment Concrete Layer Thickness (inch) CAB Layer Thickness (inch) Reinforcement ID No. Type Location PCC-Contr-IS 7 6 8 None (Control) N/A PCC-Grid-IS 9 6 8 Geogrid Middle of the Base PCC-Textile-IS 10 6 8 Geotextile Middle of the Base Note: N/A = not applicable. Table 3.3. Loading Protocol for Flexible and Rigid Pavement in LST Experiments Load Type Target Load Level (kips) Loading Duration Rest Period AC PCC AC PCC AC PCC Dynamic (0.1-sec Loading + 0.9-sec Rest Period) 9 80 Cycles 25 Cycles 5 Minutes 3 Minutes 12 100 Cycles 16 150 Cycles Static 6 5 5 Minutes 3 Minutes 30 Minutes 5 Minutes 9 12 Triaxial Laboratory Testing The research team tested two different base courses under triaxial laboratory conditions according to protocols for resilient modulus and permanent deformation testing that were developed before this project to obtain the anisotropic properties of a base course material. One
11 was a base course that was used in the LST tests at the University of Nevada, Reno, and the other was a limestone base course commonly used in Texas. Both geotextiles and geogrids were placed at three different locations within the samples, and their effects on the anisotropic properties and permanent deformation properties of the base courses were measured. The protocols that were used in these tests include measurements of the resilient modulus, anisotropic ratio, and permanent deformation with repeated loading. Finite Element Modeling The finite element program ABAQUS that was available in the computer system was used to model the effects of loading in the different tests that were performed in the LST. The computer program was capable of representing anisotropy, stress dependency, and plasticity zones (42). It also included a Goodman-type interface element that was capable of modeling the slippage and interaction between the base course and the geosynthetics (43). This program was used to match, to the extent possible, the measured results of the LST tests. The mechanical effects of geosynthetics on the properties of an unbound base course were two-fold. They affected the stiffness, anisotropic ratio, and permanent deformation of the base course as measured in the triaxial test. They also provided the membrane effect on increasing the confining pressure and reducing the vertical strain in the base course and subgrade. Although only the stiffening effect was measured in the triaxial test, both the stiffening and the membrane effects were combined in the finite element analyses. Development of ANN Models of Critical Strains and Stresses Using the finite element program that successfully matched the measurements in the LST tests, the researchers made a large number of runs covering a wide range of pavement variables. In flexible pavements, the variables included various thicknesses of the asphalt, base, and subgrade; various levels of moduli of each layer; various anisotropic ratios of the base course; and several levels of sheet stiffness and locations of geosynthetics. No ANN models of critical stresses were developed for rigid pavements because the finite element model computations, as illustrated in Chapter 4, showed that critical stresses were insensitive to the type and location of geosynthetics. In all cases for flexible pavements, one load level was used: the 9-kip load. The anisotropic ratios used in the computations spanned the range that was measured in the triaxial tests of the two base courses. In this way, the alteration of the anisotropic ratios by the embedded geosynthetics was accounted for. In each case, the critical strains and stresses in pavements were calculated. These strains or stresses were used in Pavement ME Design to predict pavement distresses and roughness. The load-related pavement distresses that were predicted in flexible pavements were roughness, rutting, and fatigue cracking. A new permanent deformation model replaced the permanent deformation model of the base course and subgrade in the Pavement ME Design software because of its superior predictions at different stress states and numbers of load repetitions in the repeated load triaxial tests. ANN models were developed for each of the critical strains and stresses in flexible pavements using the extensive computed database that was generated with the finite element program. These stresses were inputs in the new permanent
12 deformation model. When geosynthetics were used in the design of a flexible pavement, these ANN models could be used in place of the current models in the Pavement ME Design software, as indicated schematically in Figure 1.1. The material property inputs to the Composite GeosyntheticâBase Course Model developed in this project included separate properties of the unreinforced, unbound base course and the sheet stiffness of the geosynthetic. These properties could be inputs at Level 1, 2, or 3. ï· Level 1: Base course input data included the stress-dependent coefficients for both the vertical modulus and the shear modulus, the anisotropic ratio, a shear interaction coefficient between the base course and the geosynthetic, and the suction-vs.-water content characteristic curve of the base course. The permanent deformation model required two exponents in addition to the input required by the Pavement ME Design software. The geosynthetic properties included the sheet stiffness and its location within the base course. If geogrids were used, the aperture needed to be at least 1.2 times the maximum size of the aggregate of the base course. ï· Level 2: Base course input data included a modulus; a Poissonâs ratio; an anisotropic ratio of the unreinforced, unbound base course; and a shear interaction coefficient between the base course and the geosynthetic. The geosynthetic input data included the sheet stiffness and its location within the base course. ï· Level 3: Base course input data included a selection of typical base courses with tabulated material properties. A typical tabulated shear interaction coefficient was used. The geosynthetic input data included the sheet stiffness and its location within the base course. The separate properties of the base course and geosynthetic were combined to make a composite material, the properties of which were used in the ANN models to calculate the critical strains and stresses of pavements. The rules for making this conversion into a composite material are given in Chapter 4. Performance Data Collection of In-Service Pavement Sections with Geosynthetics The research team identified the in-service pavement sections with embedded geosynthetics from the Long-Term Pavement Performance (LTPP) database and other databases from local highway agencies. The collected pavement structure and performance information was listed as follows: ï· Pavement structure data, including layer thickness, construction dates, material design information, and falling weight deflectometer data. ï· Traffic data from the identified pavement sections, which should be compatible with the input of the traffic module in the Pavement ME Design software. ï· Climatic data or weather station information from the identified pavement sections. ï· Performance data from the identified pavement sections, including fatigue cracking, rutting, and international roughness index.
